Several abnormal neuropsychiatric traits are more prevalent among the nonschizophrenic biological relatives of schizophrenics than is clinical schizophrenia. Prominent among such potential expressions of "latent liability" are eye movement dysfunctions, thought disorder, and schizophrenia spectrum disorders and traits. The availability of alternative, more prevalent manifestations of the schizophrenia genotype would increase the power of linkage studies by broadening the pool of informative families. To be optimally informative for linkage studies, however, methods for accurately measuring these traits need to be established, including information about their distribution, specificity, stability, and interrelatedness. Each of these traits has been studied extensively in schizophrenics and their family members. This study addresses the co-aggregation within the same individuals of these traits and their individual and collective usefulness in discriminating 105 schizophrenics from 105 normal controls and 75 siblings of schizophrenics from 75 sibling of controls. We will use the discriminant analysis, logistic regression, multivariate analysis of variance, and normal mixture models to compare schizophrenics with normal controls and siblings of schizophrenics with siblings of normals. These data complement those from the McLean Schizophrenia Collaborative Project ("Collaborative Biological Research in Schizophrenia"; MH 31154), which is studying the co-aggregation of these traits (and others) within families of schizophrenics, but in a different family design.